This specification relates to tracking the origin of images.
A search system can access resources available on a network and maintain an index or multiple indexes of the resources. For example, the resources could include HTML pages, word processing documents, and portable document format (PDF) documents, feed sources, and media such as images and video. The index of the resources can be queried to receive the resources or information about the resources.
With respect to images, many services take a particular image and produce slightly modified or at least recompressed versions of the particular image. Such modifications change the byte fingerprint of the image so that such fingerprints cannot be used to identify the images as the same image. For example, an image search engine crawls images and produces large and small thumbnails for serving with image search results. Likewise, an Internet photo management website often produces several different versions of an uploaded image and publishes these different versions. These are just a few examples of how many variations of a source image are produced and published.